# initialisation code copied on file GEO872-session-01-solution.rmd

# packages:

## Default repository
local({r <- getOption("repos")
       r["CRAN"] <- "http://cran.r-project.org" 
       options(repos=r)
})

check_pkg <- function(x)
  {
    if (!require(x,character.only = TRUE, quietly = TRUE))
    {
      install.packages(x,dep=TRUE)
        if(!require(x,character.only = TRUE, quietly = TRUE)) stop("Package not found")
    }
}

# Call check_pkg() to install/load the required packages. 
check_pkg("sf")
check_pkg("readr") 
check_pkg("ggplot2")
check_pkg("dplyr")
check_pkg("lubridate") #for timezones
check_pkg("raster")
check_pkg("leaflet")
#check_pkg("wordcountaddin") -> ma: does not work on my computer :(

Criteria:

length: 30’000 - 50’000
Report your work in a written project report. The report has two functions:
– It shall serve you as documentation of what you did, such that at a later stage you can use the report in one of your own projects.
– It will be used to evaluate and mark your project.

Your report shall: – cover how you went about investigating your research questions. Describe your data science ideas and how you implemented your ideas. – present the results of your study and discuss them in the light of your research questions. What have you achieved and what would be further steps for future research? – report problems and limitations you encountered along the way and the solutions you chose to overcome these, be it limitations with respect to the data sources, the tools or any other source of limitation. – discuss your data science choices in the light of the theory covered in the lectures, group works, and your reading assignments.

Good example: https://fbilan.github.io/GEO880_2020/Project_GEO880_Wysling_Biland.html

Introduction

Research questions

Do the characteristics of movement trajectories differ beetween different travel modes?
Could the characteristics be used to identify travel mode?
How to identify different travel modes with movement trajectories? Could a identification model be built?
How accurate and efficient could different travel modes be identified based on the aforementioned identification model?

Background

Literature review - list of travel modes considered in our study. - characteristics specific to the mean of transport

Methodology

First the trajectories data will be prepared to be fed for machine learning: data cleaning, labelizing, values extracted.

Labels list: boat, train, etc.

Preprocessing of trajectories data

Posmo data

clean GPS data and add basic common columns attributes
labelize data
overview (here or maybe later?) …

Import, clean posmo data

#import and convert to sf object
posmo <- read_csv("./data/posmo_2023-06-14_2.csv") %>% st_as_sf(coords = c("lon_x", "lat_y"), crs=4326, remove = FALSE) |> st_transform(2056)

## Rows: 50516 Columns: 7
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr  (4): user_id, weekday, place_name, transport_mode
## dbl  (2): lon_x, lat_y
## dttm (1): datetime
## 
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

#clean and add columns
posmo$source <- "posmo"

posmo <- posmo %>% 
       rename("tmode_posmo" = "transport_mode")

posmo$tmode_manual <- NA

to_remove <- c("place_name", "lon_x", "lat_y", "user_id")
posmo <- posmo[ , !(names(posmo) %in% to_remove)]

#add East and North columns
coords <- posmo |> st_coordinates()
posmo <- posmo |> 
  mutate(E =  coords[,1], N = coords[,2])

#update timezone
posmo <- posmo %>%
  mutate(
    datetime = datetime %>% with_tz(tzone = "Europe/Zurich")
  )

# Add heights
# The height comes from the Digital height model DHM25 provided by swisstopo. The raster has been preprocessed on QGIS to be converted into the epsg 2056 (previously 21781) to save calculation time.
# The resolution is 25m. To have more precise information, we could look at the swiss3DAlti if needed.  
# Simplification: we did not consider height of soil compared to height of person or public transport.  
# As it is a relative variable to calculate slope, it does not matter.

#Extract values from rasters:
#https://gisday.wordpress.com/2014/03/24/extract-raster-values-from-points-using-r/comment-page-1/

dhm25 <- raster("./data/dhm25_raster/dhm25_2056.tif")
rasValue <- extract(dhm25, posmo)
posmo$H <- rasValue

# #Adapt column order
# col_order <- c("source", "datetime", "weekday", "tmode_posmo", "tmode_manual" ,"E", "N", "H", "timelag", "steplength_m", "speed_ms", "slope", "geometry")
# posmo <- posmo[, col_order]
# posmo

Labelize posmo data

#select segments
condition_boats <- posmo$datetime > as.POSIXct("2023-05-14 13:55:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-05-14 14:21:00", tz = "Europe/Zurich")

condition_cable_cars <- 
  posmo$datetime > as.POSIXct("2023-04-12 09:12:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 09:18:30", tz = "Europe/Zurich")

condition_ski_lifts <- 
  posmo$datetime > as.POSIXct("2023-04-12 09:23:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 09:30:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-04-12 09:33:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 09:41:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-04-12 09:51:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 9:59:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-04-12 10:15:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 10:24:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-04-12 10:51:30", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 10:57:30", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-04-12 11:00:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 11:08:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-04-12 11:21:30", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 11:29:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-04-12 11:46:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 11:51:30", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-04-12 13:49:30", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 13:54:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-04-12 13:57:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 14:05:00", tz = "Europe/Zurich")

condition_t_bars <- 
  posmo$datetime > as.POSIXct("2023-04-12 10:28:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 10:33:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-04-12 14:10:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 14:18:00", tz = "Europe/Zurich") 

condition_skis <- 
  posmo$datetime > as.POSIXct("2023-04-12 09:30:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 09:33:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-04-12 09:41:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 9:51:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-04-12 09:59:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 10:15:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-04-12 10:24:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 10:28:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-04-12 10:33:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 10:51:30", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-04-12 10:57:30", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 11:00:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-04-12 11:08:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 11:21:30", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-04-12 11:29:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 11:46:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-04-12 11:51:30", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 13:49:30", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-04-12 13:54:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 13:57:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-04-12 14:05:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 14:10:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-04-12 14:18:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 14:48:00", tz = "Europe/Zurich") 

condition_scooters <- 
  posmo$datetime > as.POSIXct("2023-05-04 07:36:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-05-04 07:42:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-05-24 22:00:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-05-24 22:10:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-05-24 19:48:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-05-24 20:00:00", tz = "Europe/Zurich")

condition_bikes <- 
  posmo$datetime > as.POSIXct("2023-05-02 08:31:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-05-02 08:36:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-05-02 11:49:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-05-02 12:04:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-05-02 17:44:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-05-02 18:04:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-05-15 15:57:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-05-15 16:32:00", tz = "Europe/Zurich") 

condition_trains <- 
  posmo$datetime > as.POSIXct("2023-03-30 17:00:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-03-30 17:52:00", tz = "Europe/Zurich") | 
  posmo$datetime > as.POSIXct("2023-05-26 18:10:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-05-26 19:15:00", tz = "Europe/Zurich") | 
  posmo$datetime > as.POSIXct("2023-05-09 8:40:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-05-09 10:55:00", tz = "Europe/Zurich") | 
  posmo$datetime > as.POSIXct("2023-04-28 12:20:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-28 12:50:00", tz = "Europe/Zurich") | 
  posmo$datetime > as.POSIXct("2023-06-09 19:30:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-09 21:30:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-04-03 22:15:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-04 00:22:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-06-09 12:23:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-09 12:48:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-06-11 02:30:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-11 02:45:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-06-12 07:00:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-12 07:53:00", tz = "Europe/Zurich") | 
  posmo$datetime > as.POSIXct("2023-06-14 17:25:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-14 17:42:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-06-06 09:05:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-06 11:00:00", tz = "Europe/Zurich")

condition_trams <- 
  posmo$datetime > as.POSIXct("2023-04-16 15:30:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-16 15:37:00", tz = "Europe/Zurich") | 
  posmo$datetime > as.POSIXct("2023-06-09 17:17:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-09 17:22:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-06-09 13:40:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-09 13:54:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-06-06 11:09:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-06 11:23:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-06-05 18:49:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-05 19:03:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-06-05 23:33:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-05 23:40:00", tz = "Europe/Zurich")

condition_bus <- 
  posmo$datetime > as.POSIXct("2023-06-09 17:25:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-09 17:31:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-05-24 19:32:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-05-24 19:49:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-05-24 22:10:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-05-24 22:22:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-06-07 07:53:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-07 08:12:00", tz = "Europe/Zurich") | posmo$datetime > as.POSIXct("2023-06-07 08:35:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-07 08:45:00", tz = "Europe/Zurich") | posmo$datetime > as.POSIXct("2023-06-07 19:23:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-07 19:37:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-06-06 20:39:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-06 20:44:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-06-06 21:03:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-06 21:17:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-06-06 08:15:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-06 08:25:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-06-06 08:28:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-06 08:35:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-06-05 18:37:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-05 18:46:00", tz = "Europe/Zurich") 

condition_walks <- 
  posmo$datetime > as.POSIXct("2023-04-16 15:30:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-16 15:37:00", tz = "Europe/Zurich") | 
  posmo$datetime > as.POSIXct("2023-04-16 15:44:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-16 15:56:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-05-13 13:29:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-05-13 15:01:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-05-13 16:40:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-05-13 22:01:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-05-13 22:04:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-05-13 22:29:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-05-20 11:54:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-05-20 22:35:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-05-22 11:00:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-05-22 11:42:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-05-29 16:18:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-05-29 16:46:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-06-09 11:30:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-09 12:20:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-06-11 02:46:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-11 03:04:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-06-12 07:53:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-12 08:07:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-06-06 08:11:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-06 08:15:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-06-05 23:40:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-05 23:57:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-04-12 09:00:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 09:12:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-04-12 18:40:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-12 18:50:00", tz = "Europe/Zurich")

condition_cars <- 
  posmo$datetime > as.POSIXct("2023-06-11 15:00:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-11 17:00:00", tz = "Europe/Zurich") | 
  posmo$datetime > as.POSIXct("2023-04-24 14:00:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-04-24 17:30:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-06-10 13:40:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-10 17:33:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-06-09 21:45:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-09 22:30:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-06-14 07:10:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-14 08:00:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-06-14 17:44:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-14 18:10:00", tz = "Europe/Zurich") |
  posmo$datetime > as.POSIXct("2023-06-13 18:00:00", tz = "Europe/Zurich") & posmo$datetime < as.POSIXct("2023-06-13 19:15:00", tz = "Europe/Zurich") 

#final labelizing
posmo <- posmo %>%
  mutate(tmode_manual = case_when(condition_walks ~ 'walk',
                                  condition_boats ~ 'boat',
                                  condition_trains ~ 'train',
                                  condition_bikes ~ 'bike',
                                  condition_scooters ~ 'kick-scooter',
                                  condition_cars ~ 'car',
                                  condition_bus ~ 'bus',
                                  condition_trams ~ 'tram',
                                  condition_t_bars ~ 't_bar',
                                  condition_cable_cars ~ 'cable_car',
                                  condition_ski_lifts ~ 'ski_lift',
                                  condition_skis ~ 'ski'
                                  ))

GPS data

clean GPS data and have basic common columns attributes
labelize data
overview …

Merge Posmo and GPS tracker

Compute new variables: timelag, steplength, speed, acceleration, slope, etc.
remove statics (ex.3)
Add segment IDs (ex.3 - task 4)

#here merge with gps tracker
mvmt_data <- posmo

Overview of Posmo and GPS tracker datasets:

Map general overview - two users:

#map overview
mvmt_data_wgs <- mvmt_data %>% st_transform(crs = 4326)

m <- leaflet() %>%
  addTiles() %>%  # Add default OpenStreetMap map tiles
  addCircleMarkers(data = mvmt_data_wgs,
                   #opacity = 0.3, 
                   radius = 0.2,
                   popup = mvmt_data_wgs$datetime,
                   color = 'blue') %>%
  addLegend(position = 'topright', 
            colors = c('blue'), 
            labels = c('points'),
            title = 'All points from GPS tracker and posmo')

# Plot the leaflet object m
m

Map general overview - posmo user:
…

Map general overview - GPS tracker user:
…

Map general overview - per region (?):
…

Compute new variables

Here we will compute all necessary values to determine mean of transport:

speed (steplength + timelag)
acceleration
sinuosity
slope
(elevation speed?)
…
region (mainly for visualisation)

#compute new values
#timelag
mvmt_data <- mutate(mvmt_data, timelag = as.numeric(difftime(lead(datetime), datetime)))

#steplength
mvmt_data <- mvmt_data %>% mutate(steplength_m = sqrt((E-lead(E))^2 + (N-lead(N))^2))

#speed
mvmt_data <- mvmt_data %>% mutate(speed_ms = steplength_m/timelag)

# Add slope
mvmt_data <- mvmt_data %>% mutate(slope = (lead(H) - H)/steplength_m * 100)

# add regions
#Zurich
zh_ll <- c(2674900, 1244000)
zh_ur <- c(2692100, 1260300)

#Waadt
wt_ll <- c(2480000, 1127600)
wt_ur <- c(2575200, 1195100)

#wallis
ws_ll <- c(2538000, 1076875)
ws_ur <- c(2661500, 1153125)

mvmt_data$region <- NA
mvmt_data <- mvmt_data |> 
  mutate(
    region = case_when(
      E > zh_ll[1] & E < zh_ur[1] & N > zh_ll[2] & N < zh_ur[2] ~ 'Zurich', 
      E > wt_ll[1] & E < wt_ur[1] & N > wt_ll[2] & N < wt_ur[2] ~ 'Waadt',
      E > ws_ll[1] & E < ws_ur[1] & N > ws_ll[2] & N < ws_ur[2] ~ 'Wallis'
      )
    )

Segmentize and deal with statics points

…

Annotate trajectories with environment datasets

datasets descriptions
annotate trajectories

Summarize characteristics of the means of transport

Compute mean variables per segment (ex.2 task 5 (rolling windows)) -> summarize the variables by mean of transports
general overview for each mean of transports

Summary all segments from the mean of transport:

#count of segments:
tmode_count <- mvmt_data %>% group_by(tmode_manual) %>% count() %>% arrange(desc(n))

knitr::kable(tmode_count %>% st_drop_geometry(),
caption = "Number of segments labelized per mode of transport")

Number of segments labelized per mode of transport
tmode_manual	n
NA	37574
train	4875
car	2401
walk	2357
ski	814
bus	690
tram	502
bike	498
ski_lift	477
kick-scooter	116
boat	85
t_bar	84
cable_car	43

#maps per mean of transport?:

mvmt_data_filter <- mvmt_data %>% filter(tmode_manual == "train") #%>% filter(!is.na(tmode_manual)) #%>% filter(region == "Zurich") 

mvmt_data_filter_wgs <- mvmt_data_filter %>% st_transform(crs = 4326)

m <- leaflet() %>%
  addTiles() %>%  # Add default OpenStreetMap map tiles
  addCircleMarkers(data = mvmt_data_filter_wgs,
                   #opacity = 0.3, 
                   radius = 0.2,
                   popup = mvmt_data_filter_wgs$datetime,
                   color = 'blue') %>%
  addLegend(position = 'topright', 
            colors = c('blue'), 
            labels = c('points'),
            title = 'All segments labelized as (selected)')

# Plot the leaflet object m
m

Machine Learning

…

Results

…

Discussion

Answer the 4 RQ

Problems encoutered / challenges

posmo map matching

Limitations

…

Conclusion

…

GEO880-G3-Report

Martine Besse & Yelu He

2023-06-03